K3 crystal filter offset/bandwidth/gain setting guidelines

[wayne burdick]

Here are my thoughts on K3 crystal filter setup, inspired by a recent
question posted to the Elecraft K3 Yahoo group by Eric Scace, K3NA.


Offset (FLx FRQ)
----------------

There are many considerations that go into the positioning of the K3's
crystal filters with respect to the I.F. center frequency. We shift
them upward if the filter is too wide to be centered at Fc/2 + 200 Hz,
thus preserving the lower edge of the filter around 200 Hz. This
usually happens only in CW mode, typically with lower sidetone pitches.
The result is optimization for opposite-sideband rejection. You can
argue for a different approach, but the K3's success in CW DXing and
contesting suggests that this approach is just as valid as any other.

I don't recommend trying to fool the firmware by adjusting the crystal
filter offsets; I'd use the marked values. Changing them is likely to
cause unwanted side-effects, since the filter passband is inverted for
complimentary modes (CW/CW REV, LSB/USB, DATA/DATA-REV). The exception
to this rule is when fine-tuning the offsets of 5-pole filters on the
main and sub-receivers in order to provide best performance in
diversity mode. These adjustments will rarely be larger than +/- 20 Hz
anyway.

Bandwidth (FLx BW)
------------------

It's OK (but not necessary) to fudge the bandwidth of specific filters
(FLx BW). For example, INRAD's 8-pole "400 Hz" filter can be declared
as 0.4, 0.45, or 0.5 kHz, depending on where you want this filter to be
kicked in as WIDTH is rotated. The audible effect is subtle. Ed Muns,
W0YK, has described this technique in detail elsewhere.

Gain (FLx GN)
-------------

Regarding filter loss compensation: I recommend simply sticking with
what's on page 46 of the owner's manual (1 to 2 dB compensation for
400-500 Hz filters, 3 to 4 dB compensation for 200-250 Hz filters, and
0 for all others) unless you have a very specialized application and
lots of time on your hands.

There are three reasons for this. (1) The *perceived* loss of a crystal
filter is a function of both actual loss and S/N ratio. If a narrower
filter has a little more loss, you may not notice it. (2) Most signals
are flattened slightly by AGC anyway. (3) Adding a lot of gain to
filters can again cause side-effects, such as interaction with RX EQ
settings and variations in gain between main and sub if the filter
complements are different (e.g., during diversity use).

In fact we recently issued new guidelines for factory-assembled K3s:
all crystal filters' FLx GN numbers will be to 0 except for 400-500 Hz
(1 dB) and 200-250 Hz (2 dB). Customers are free to optimize these
further, of course -- but most won't need to.

73,
Wayne
N6KR

---
http://www.elecraft.com

______________________________________________________________
Elecraft mailing list
Home: http://mailman.qth.net/mailman/listinfo/elecraft
Help: http://mailman.qth.net/mmfaq.htm
Post: mailto:[hidden email]

This list hosted by: http://www.qsl.net
Please help support this email list: http://www.qsl.net/donate.html

[AA4BQ]

Interesting suggestions. I would like to know the specific effects of shifting from the marked values. Here's my experience with this topic:

My K3 build went as advertised. I followed the instructions to the letter including recording all the marked values of the 5-pole, 500 Hz and 200 Hz filter offsets, (-.77 and -.91, respectively). Both receivers had the same values. (I ordered the 40 Hz matching option, not really understanding it.) In any case, this part of the build process was a mystery until well after getting the K3 up and running. I have literally spent the better part of a week setting and experimenting with the options in the CONFIG menu (excluding experimentation with the TECH settings (TECH MD = OFF).  In the process, I have fine tuned the performance of the K3 to where I believe is about as good as it can get. There is one exception. When selecting the 200 Hz filter (-.91) in both receivers, and crank the bandpass down to 200 as well, received signals virtually disappear. No matter how carefully I tune (I have the P3 also), the audio is non-existent (practically speaking). Cranking up the gain to 8db on each filter does little to no good. The only way to begin getting any performance at this bandpass is to modify the filter offsets. The peak performance comes at a setting of -.71 for both filters. Now, I am not entertaining the notion of having had filters mis-marked. I first assume that I am missing something. But, believe you me, I have tried EVERYTHING I can think of. Of course it seems to me that modifying the setting from factory markings (-.91) is somehow fooling the rig, possibly degrading it's performance somehow. Having said that, it performs beautifully from the perspective of audio bandbass performance with the offsets modified in this way. However, I have not experimented with adjacent strong signals such that the roofing filters show their true colors, so to speak, set up in this way. During ARRL Field Day, there were many opportunities for me to exercise and test the K3's performance set both ways (-.91 and -.71). It is my opinion that I would not go back to -.91 based on my listening/operating satisfaction. Unless there is some other setting that I have overlooked that is preventing these 200 Hz filters from working when the bandpass is set down to 200 hz, then I must stay with this. Otherwise, there is no advantage over the 400 hz bandbass. Where am I going wrong here?

Perhaps my experience is helpful to someone, somehow. Similar experiences would enlighten the conversation I think. A solution would put it to bed. Thanks. I hope this helps.